Event boundaries drive norepinephrine release and distinctive neural representations of space in the rodent hippocampus

Abstract

Episodic memories are temporally segmented around event boundaries that tend to coincide with moments of environmental change. During these times, the state of the brain should change rapidly, or reset, to ensure that the information encountered before and after an event boundary is encoded in different neuronal populations. Norepinephrine (NE) is thought to facilitate this network reorganization. However, it is unknown whether event boundaries drive NE release in the hippocampus and, if so, how NE release relates to changes in hippocampal firing patterns. The advent of the new GRAB_NE sensor now allows for the measurement of NE binding with sub-second resolution. Using this tool in mice, we tested whether NE is released into the dorsal hippocampus during event boundaries defined by unexpected transitions between spatial contexts and presentations of novel objections. We found that NE binding dynamics were well explained by the time elapsed after each of these environmental changes, and were not related to conditioned behaviors, exploratory bouts of movement, or reward. Familiarity with a spatial context accelerated the rate in which phasic NE binding decayed to baseline. Knowing when NE is elevated, we tested how hippocampal coding of space differs during these moments. Immediately after context transitions we observed relatively unique patterns of neural spiking which settled into a modal state at a similar rate in which NE returned to baseline. These results are consistent with a model wherein NE release drives hippocampal representations away from a steady-state attractor. We hypothesize that the distinctive neural codes observed after each event boundary may facilitate long-term memory and contribute to the neural basis for the primacy effect.

Figure 1.

Time from context transition controls Signal_NE when mice are moved to novel arenas. A) Histological confirmation of GRAB_NE expression (GFP) and fiber placement over dorsal CA1. B) Schematic of experimental timeline. C) Example session showing increases in Signal_NE around each context and homecage (HC) transition. D) Mean Signal_NE measured across all transitions (black) and cross-validated prediction from the saturated model (red) or a reduced model lacking terms related to time from transfer (blue). E) Change in CVMSE due to removal of various potential behavioral variables. Only removal of the terms related to time from transition significantly decreased model performance (t(7) = 3.30, p = 0.01).

Figure 2.

Time from context transition controls Signal_NE when mice are moved to a linear track. A) Example session showing Signal_NE (black) aligned with acceleration (red) and reward delivery (·). Vertical gray lines show that local peaks in Signal_NE do not align to bouts of acceleration nor reward timing. Shaded area shows last 60s before removing from track during which Signal_NE was not modeled. B) Mean Signal_NE measured across all linear track transitions (black) and cross-validated prediction from the saturated model (red). C) Change in CVMSE due to removal of various potential behavioral variables. Only removal of the terms related to time from transition significantly decreased model performance (t(7) = 7.20,p = 0.0008).

Figure 3.

Time from object introduction controls Signal_NE A) Photographs of five novel objects presented to the mouse. B) Example session showing Signal_NE (black) aligned object introduction (dashed line) and object sampling (·). C) Mean Signal_NE measured across all object presentations (black) and cross-validated prediction from the saturated model (red). C) Change in CVMSE due to removal of various potential behavioral variables. Only removal of the terms related to time from object introduction significantly decreased model performance (t(5) = 3.54, p =0.017).

Figure 4.

Novel objects do not affect NE dynamics after transfer to a familiar linear track. A) Mean Signal_NE across experimental sessions when the track was baited with a novel object (black); control sessions were run without new objects (red). B) Estimated τ describing Signal_NE decay after moving to the linear track did not change in the presence of a novel object (t(4) = 1.47, p = 0.22). C) Change in CVMSE due to removal of various potential behavioral variables. Only removal of the terms related to time from linear track transfer significantly decreased model performance (t(5) = 3.22, p = 0.03).

Figure 5.

Experience accelerates Signal_NE decay after context transition. A) Mean Signal_NE plotted as a function of time from context transition (dashed line) and color coded by number of days of experience. Black trace shows Signal_NE recorded after transitioning back to the home cage (HC). B) Estimated Signal_NE derived from the saturated model. C) Parameter estimates for the magnitude (β) and decay rate (τ) of Signal_NE after context transition color-coded by days of experience. D) Decay rate (τ) after transfer to the arena hastens over days of exposure (mixed-effect linear model; t(73) = 2.31, p = 0.02) and is most rapid during transfer to the HC (Day N vs HC, all p ≤0.01).

Figure 6.

Signal_NE is depressed relative to baseline after periods of sustained elevation. A) Mean Signal_NE recorded after moving mice back to the home cage from the arena (red) or the linear track (black). B) Same data as Figure 5C with the addition of parameter estimates for the behavior of Signal_NE after transition to home cage from the linear track. C) The decrease in Signal_NE was significantly larger after transitioning mice to the home cage from the linear track as compared to from the novel arenas (t(5) = 3.74, p = 0.005)

Figure 7.

CA1 spatial code takes minutes to stabilize after context transition in novel and familiar spaces. A) Example UMAP embedding of population firing rate vectors (100-ms), color-coded by where the mouse was physically located on a linear track when the data was recorded. B) Same embedding color coded by time from context transfer. C) Representational similarity (Pearson R) of the observed population firing rate vector at each moment in a novel environment relative to the mean of the next 3 most similar vectors recorded in the same location. D) Same as Panel C recorded in a familiar environment. E) In a novel environment, the patterns recorded in the first minute were less correlated than those observed 10 minutes into the session (t(7) = 8.05, p = 0.00009) F) Same as Panel E recorded in a familiar environment (t(7) = 8.20, p = 0.00008). G) Initial representations were more correlated to their nearest neighbors in a familiar environment as compared to those recorded in a novel environment (t(7) = 7.58, p = 0.0001).

Figure 8.

Moments immediately after transition are not preferentially replayed. A) Percentage of ripples recorded before (black) and after (red) experiencing a novel environment that showed significant reactivation of each moment after transition. Dashed line shows false positive (FP) rate. B) Moments recorded 10–11 minutes after novel context transition were more likely to be reactivated than those recorded 0–1 minutes after transition (t(7) = 2.46, p = 0.04). C) Same as Panel C showing reactivation rates as a function of time after transition to a familiar environment. D) There is no difference in reactivation rate for early vs late moment in a familiar environment (t(7) = 0.40, p = 0.70).